Version information

2.1.0

• OPC UA Connector support

• Support for SIMATIC S7 Connector and SIMATIC S7+ Connector: additional data types

• Support for 3rd party Vision Connectors

• Writing data to PLC

• Shared memory support between Docker containers and Linux processes with RIB

• Updated user interface

• Support for the Chinese language

2.0.0

• Execution of AI models on GPU-accelerated hardware

• Support for additional data types: ByteArray, SIntArray, Int8Array, Int16Array, RealArray (uses float32), LRealArray (uses float64)

• Support for batch processing

• Support of multiple Python versions (3.8.19, 3.10.14)

• Support for retention of pipeline output data via the Databus

• Simpler installation with the Edge application AI Model Manager

• Adjustments for standardized image transport

• Updated user interface

1.6.0

• Option for enabling automatic restart of the app, which continues execution of the AI model after multiple reboots of the Edge device.

• Support for the mass mapping of several identical inputs

• Advanced filter options for log messages

• Support for the ImageSet type

• Support for wildcarded topics

• Improvements to the user interface

• Permanent storage for the pipelines

1.5.0

• Support for execution of multiple pipelines (up to 3)

• Versioning of pipeline packages based on the Package ID (UUID)

• Option to send data to an external ZMQ server and receive data from an external ZMQ

• Improvements to the user interface

1.4.0

• Simplified setup of the data connection

• Immediate log deletion

• Writing the data (while using the multi-part message format) to external ZMQ

• Removal of the standard connection

• Display of data sources in the pipeline visualization

• Generation of more than one data connection with the same data connection type and the same payload (although with different login information)

• Improvements in debugging and logging

• Improvements to the user interface

1.3.0

• Simultaneous parallel step execution in the same pipeline

• Support of Openness API

• Inactivation of the pipeline

• Improvements to the user interface

1.2.0

• Updating the model/pipeline parameters on the user interface of the AI Inference Server before execution of the pipeline

• Delta upload of the configuration package

• Support of IE Vision Connector via Databus (MQTT) or directly via ZMQ

• Support of application upgrade via Industrial Edge Management

1.1.0

• Modification of the user interface

• Support of multiple connectors

• Automatic start function after restart

• The AI model can retrieve data directly from various Edge connectors

• Fine tuning for "internal" queue (ZMQ) per pipeline with App Config

• Logging improvements

1.0.1

AI Inference Server version 1.0.1 is being launched on the market.